摘要
针对钢管修磨控制系统中存在的常见故障,构造了神经网络信息融合中心。对来自多传感器的残差信号进行了预处理和离散小波变换,提取其细节系数作为神经网络的故障特征向量,使用改进BP算法对神经网络分类器训练以进行相应的故障模式识别。仿真结果表明,基于神经网络的信息融合技术用于控制系统的故障诊断是可行的和有效的。
In view of the fault that often appears in nonlinear system such as fixing s teel pipes,the information fusion pattern based neural network is constructed.The residual signal from multisensor is preprocessed an d dis-crete wavelet transformated,the detail coefficients are obtained as fa ult character vectors,neural netwo rk clas-sifier is trained to complete fault p attern recognition by modified BP algorithm.Simulation results reveal that the application of fault diagnosis of co ntrol system based on neural network information fusion is practical and effective.
出处
《安徽工业大学学报（自然科学版）》
CAS
2004年第2期171-174,共4页
Journal of Anhui University of Technology（Natural Science）
关键词
故障诊断
信息融合
神经网络
小波变换
fault diagnosis
information fusion
neural networks
wavelet transformation
